A current challenge in microbial pathogenesis is to identify biological control agents that may prevent and/or limit host invasion by microbial pathogens. In natura, hosts are often infected by multiple pathogens. However, most of the current studies have been performed under laboratory controlled conditions and by taking into account the interaction between a single commensal species and a single pathogenic species. The next step is therefore to explore the relationships between host-microbial communities (microbiota) and microbial members with potential pathogenic behavior (pathobiota) in a realistic ecological context. In the present study, we investigated such relationships within root-associated and leaf-associated bacterial communities of 163 ecologically contrasted Arabidopsis thaliana populations sampled across two seasons in southwest of France. In agreement with the theory of the invasion paradox, we observed a significant humped-back relationship between microbiota and pathobiota α-diversity that was robust between both seasons and plant organs. In most populations, we also observed a strong dynamics of microbiota composition between seasons. Accordingly, the potential pathobiota composition was explained by combinations of season-specific microbiota operational taxonomic units. This result suggests that the potential biomarkers controlling pathogen's invasion are highly dynamic.

14A current challenge in microbial pathogenesis is to identify biological control agents that may 15 prevent and/or limit host invasion by microbial pathogens. In natura, hosts are often infected we observed a significant humped-back relationship between microbiota and pathobiota α-25 diversity that was robust between both seasons and plant organs. In most populations, we also 26 observed a strong dynamics of microbiota composition between seasons. Accordingly, the 27 potential pathobiota composition was explained by combinations of season-specific 28 microbiota OTUs. This result suggests that the potential biomarkers controlling pathogen's 29 invasion are highly dynamic.

Phytopathogens are a continuous threat for global food production and security. Emergence or re-emergence of plant pathogens is highly dependent on the environmental conditions affecting pathogen spread and survival. Under climate change, a geographic expansion of pathogen distribution poleward has been observed, potentially resulting in disease outbreaks on crops and wild plants. Therefore, estimating the adaptive potential of plants to novel epidemics and describing its underlying genetic architecture, is a primary need to propose agricultural management strategies reducing pathogen outbreaks, as well as to breed novel plant cultivars adapted to pathogens that might spread in novel habitats as a consequence of climate change. To address this challenge, we inoculated Pseudomonas syringae strains isolated from Arabidopsis thaliana populations located in south-west of France on the highly genetically polymorphic TOU-A A. thaliana population located east-central France. While no adaptive potential was identified in response to most P. syringae strains, the TOU-A population displays a variable disease response to the P. syringae strain JACO-CL belonging to the phylogroup 7 (PG7). This strain carried a reduced T3SS characteristic of the PG7 as well as flexible genomic traits and potential novel effectors. GWA mapping on 192 TOU-A accessions inoculated with JACO-CL revealed a polygenic architecture. The main QTL region encompasses two R genes and the AT5G18310 gene encoding for ubiquitin hydrolase, a target of the AvrRpt2 P. syringae effector. Altogether, our results pave the way for a better understanding of the genetic and molecular basis of the adaptive potential in an ecologically relevant A. thaliana - P. syringae pathosystem.